Neonatal Intensive Care Units (NICUs) have proven remarkably effective in lowering neonatal mortality and morbidity. Yet, marked variations in survival and outcome among institutions have been demonstrated when adjusted for birthweight, sex, and race. Not only are these disparities significant clinically, but they may also have serious public health and even economic impact in view of the extremely high most of NICU care. It is essential to develop a severity of illness index in order to permit accurate comparative studies of NICU care between hospitals, across regions, and over time. Severity scales have been developed for adult and pediatric intensive care which rely on the degree on physiologic abnormality of the individual patients. These scores tend to be robust in the face of variations in therapeutic style or diagnostic conventions. Such scales have not yet been attempted for neonatal intensive care, in part because of the availability of a strong predictor, birthweight. We propose to develop such a physiology based severity of illness index for neonatal intensive care over a two year period within our three- NICU system. The research strategy will be to develop an instrument using an expert panel and nominal group techniques. Score items will include basic physical and laboratory measurements. After testing reliability, it will be applied prospectively for one year in our three NICUs. Validity will be assessed by the ability to predict mortality and adverse outcomes. It will also be compared against other markers of severity including therapeutic intensity, nursing acuity, complications of intensive care (nosocomial sepsis), and hospital charges. Further refinement will come through multivariate stepwise modeling. The success of the index will be assessed by it ability to explain the variance remaining after the known major risk factors (birthweight, race, sex, plurality) have been entered. The expert panel will then reconvene to evaluate the resulting scale and recommend final revisions.
